A Recurrent Neural Network for Sentiment Quantification
This work addresses the problem of accurate sentiment prevalence prediction in text for applications like opinion mining, representing an incremental advancement in quantification methods.
The authors tackled sentiment quantification by proposing QuaNet, a recurrent neural network that learns quantification embeddings from classification predictions and refines them with simple methods, achieving substantial performance improvements over state-of-the-art baselines.
Quantification is a supervised learning task that consists in predicting, given a set of classes C and a set D of unlabelled items, the prevalence (or relative frequency) p(c|D) of each class c in C. Quantification can in principle be solved by classifying all the unlabelled items and counting how many of them have been attributed to each class. However, this "classify and count" approach has been shown to yield suboptimal quantification accuracy; this has established quantification as a task of its own, and given rise to a number of methods specifically devised for it. We propose a recurrent neural network architecture for quantification (that we call QuaNet) that observes the classification predictions to learn higher-order "quantification embeddings", which are then refined by incorporating quantification predictions of simple classify-and-count-like methods. We test {QuaNet on sentiment quantification on text, showing that it substantially outperforms several state-of-the-art baselines.